Skip to main content
. 2022 Jun 20;25(8):104644. doi: 10.1016/j.isci.2022.104644

Figure 3.

Figure 3

This figure shows the architecture of the model used for the CNN

Three similar blocks of layers with an increasing number of output filters are stacked consecutively. Finally, the output of the last block is fed into a fully connected layer with output neurons for each of the predicted motility characteristics percentage of immotile sperm, percentage of progressive motility, and percentage of non-progressive motility.